The goal of this research is to develop an automated multispectral imaging system to read protein arrays based on quantum dot encoded microparticles. Quantum dots, QDs, have recently been embedded in microparticles to generate large numbers of unique fluorescent codes. QDs are an ideal means of fluorescently encoding microparticles since they can be synthesized with emission peaks across the entire visible spectrum, yet they all have a common near-UV excitation wavelength. Tunable filters have recently been developed with the spectral and spatial resolutions needed to image 1 um diameter particles. The combination of QD encoded particles and a tunable filter makes possible novel approaches to multiplexing which could greatly increase the number of identifiable codes, and at the same time, simplify particles synthesis. Our approach is to create a binary coding scheme where each species of QD is either present or absent from the particle. By decreasing the separation between emission peaks to approximately 10 nm, theoretically it will be possible to generate more than 10[6] unique codes. Specifically, we propose to: 1) Install a tunable filter in our automated imaging system to enable collection of spectral data, 2) Write software to automate spectral image acquisition, 3) Develop algorithms for particle segmentation and classification. 4) Determine optimum ranges for QD intensities and separation of emission maxima, 5) Extend acquisition, segmentation, and classification algorithms to sedimented particles, and 6) Demonstrate immobilized and sedimented encoded particle technologies with model assay systems and evaluate assay performance. Development and testing will be greatly facilitated by incorporating the QD-coded particles into our immobilized particle arrays. The QD-encoded particles will be arrayed on hydrogel-coated slides to provide training sets with well-controlled particle densities on optically flat surfaces. They will be used to evaluate the filter performance and develop image processing algorithms. In the latter phases of our proposed research, we plan on extending the method to particles sedimented in microtiter wells and demonstrate the feasibility of detecting proteins bound to QD-encoded particles.